congbrian
commented
2 weeks ago We need more time to figure out how to do this from an architectural standpoint. The feature works as is but may not make actual sense given the deployment architecture -- currently we do not have a way to keep pythonic objects in context and pass them through our architecture, so it may not actually make sense to handle this at the API level.
Status: Backlog 10/9/2024
User Story:
As a user, I would like to be able to select not only one but multiple compatible datasets rather than a single one.
Persona:
Denzel would like to append a second dataset to his first dataset in order to retrain his model on updated information.
Feature:
Provide the possibility to mix and match datasets for a user.
Business Value:
Users will have a wider range of dataset options and combinations they can choose.
Tasks
Design an algorithm which will allow a user to concatenate a second dataset to the existing one that the api has already been connected to.
Acceptance Criteria
Python script should handle concatenation and model training smoothly after adding a dataset that _has the same labels at minimum to the old dataset
Acceptance Tests
Run model api class and select a dataset, then use concat function to concatenate a second dataset to the first one, then train a model on the combined dataset.